Abstract
Convolution neural network has important applications in the field of image recognition and retrieval, face recognition and object detection in deep learning. In the training of convolution neural network, 2D convolution, spatial pooling, linear mapping and other operations of forward propagation will have a huge computational complexity. In this paper, an effective optimization technique is proposed to map the convolutional neural network to the digital processor DSP. These technologies include: fixed-point conversion, data reorganization, weight deployment and LUT (look-up table). These technologies enable us to optimize the use of resources on the C66x DSP. The experiment is carried out on Texas Instruments C6678 development board, and the optimization technique proposed in this paper can be applied to multiple open-source network topologies.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
LeCun, Y., Bottou, L., Bengio, Y., Haffner, P.: Gradient-based learning applied to document recognition. In: Proceedings of the IEEE (1998)
LeCun, Y., et al.: CNP: An FPGA-based Processor for Convolutional Networks (2009)
A 240 G-ops/s Mobile Coprocessor for Deep Neural Networks (2013)
Putnam, A., et al.: A reconfigurable fabric for accelerating large-scale datacenter services. In: International Symposium on Computer Architecture (2014)
Jagath, A.: FPGA Implementations of Neural Networks. Springer (2006)
Optimizing FPGA-based Accelerator Design for Deep Convolutional Neural Networks. ACM 978-1-4503-3315-3/15/02 (2015)
Accelerating Deep Convolutional Neural Networks Using Specialized Hardware (2015)
Krizhevsky, A., Sutskever, I., Hinton, G.E.: Imagenet classification with deep convolutional neural networks. In: NIPS (2012)
Cong, J., Xiao, B.: Minimizing computation in convolutional neural networks. In: ICANN (2014)
Courbariaux, M., David, J.-P.: Training deep neural networks with low precision multiplications (2015)
Gupta, S., Agrawal, A., Gopalakrishnan, K.: Deep Learing with Limited Numberical Precision (2015)
Vanhoucke, V., Senior, A.: Vanhoucke improveing the speed of neural network on CPUs (2011)
Chen, T., Du, Z., Sun, N.: A Small-Footprint High-Throughput Accelerator for Ubiquitous Machine-Learning
“2020 Roadmap, Rev. 1”, European New Car Assessment Programme (March 2015)
Krizhevsky, A., Sutskever, I., Hinton, G.E.: ImageNetClassification with Deep Convolutional Neural Network. NIPS (2012)
TDA3x SoC Processors for Advanced Driver Assist Systems (ADAS) Technical Brief. Texas Instruments Inc. (2014)
Sermanet, P., LeCun, Y.: Traffic sign recognition with multi-scale Convolutional Networks. In: International Joint Conference on Neural Networks (IJCNN) (2011)
Szegedy, C., Liu, W., Jia, Y., et al.: Going Deeper with Convolutions. In: IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (2015)
“TMS320C66x DSP: User Guide”, SPRUGW0C. Texas InstrumentsInc. (2013)
Audhkhasi, K., Osoba, O., Kosko, B.: Noise benefits in backpropagation and deep bidirectional pre-training. In: 2013 International Joint Conference on Neural Networks (IJCNN), pp. 1–8. IEEE (2013)
Baboulin, M., Buttari, A., Dongarra, J., Kurzak, J., Langou, J., Langou, J., Luszczek, P., Tomov, S.: Accelerating scientific computations with mixed precision algorithms. Comput. Phys. Commun. 180(12), 2526–2533 (2009)
Acknowledgements
The authors of this paper are members of Shanghai Engineering Research Center of Intelligent Video Surveillance. In part by Technology Research Program of Ministry of Public Security of China under Grant 2015JSYJB26.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2018 Springer International Publishing AG
About this paper
Cite this paper
Cai, Y., Liang, C., Tang, Z., Li, H. (2018). Optimization Technology of CNN Based on DSP. In: Abawajy, J., Choo, KK., Islam, R. (eds) International Conference on Applications and Techniques in Cyber Security and Intelligence. ATCI 2017. Advances in Intelligent Systems and Computing, vol 580. Edizioni della Normale, Cham. https://doi.org/10.1007/978-3-319-67071-3_9
Download citation
DOI: https://doi.org/10.1007/978-3-319-67071-3_9
Published:
Publisher Name: Edizioni della Normale, Cham
Print ISBN: 978-3-319-67070-6
Online ISBN: 978-3-319-67071-3
eBook Packages: EngineeringEngineering (R0)